The risk factors and frequency of death in childhood epilepsy were assessed in a population-based cohort study at Dalhousie University, Halifax, Nova Scotia, Canada. Of 692 children who developed epilepsy during 1977-85, 26 (3.8%) had died by 1999. Frequency of death was 5.3 times higher than in the reference population in the 1980s and 8.8 times higher in the 1990s. The mortality at 20 years after onset was 6.1% compared to 0.88% in the reference population matched for age and sex. Onset of seizures in the first year carried a significantly increased risk of mortality compared to later onset epilepsy. Of 85 children with secondary generalized epilepsy, 15% died, compared to 2% of 510 with partial and primary generalized epilepsy, and 1% of 97 with absence epilepsy. Severe disorders associated with functional neurological deficit were associated with 22 times greater risk of death and were the cause of death in 22 children at a mean age of 10 years (range 1-29), independent of sex, age, and epilepsy type. Aspiration with pneumonia was the cause of death in 14, infection in 3, shunt malfunction 1, pulmonary embolism 1, congestive heart failure 1, gastroesophageal reflux 1, and status epilepticus in 1. The rate of death in patients without severe neurological deficit was 0.8% (of 591), similar to that in the reference population. Four deaths were unexpected, all in young adults, aged 18-30 years, without severe neurological deficits. Suicide was the cause in 2, homicide in 1, and a sudden unexpected death, unexplained at autopsy, in a 21-year-old woman with tuberous sclerosis who had several seizures during sleep. Functional neurological deficit was the only independent determinant of mortality. 
COMMENT. Children with epilepsy have more than five times the risk of dying in the first 20 years after onset. Most deaths are related to comorbid severe neurological disorders associated with functional neurological deficit and not to the epilepsy per se. Camfield and associates have demonstrated that death from seizures themselves is uncommon, a finding that should help to allay the fear often expressed by parents of a child with a first convulsion.
Camfields’ findings are similar to those of Callenbach PMC et al who reported a mortality rate of 3.8/1000 person-years, seven times higher than expected in a cohort of 472 children in the Netherlands. None of 328 children with nonsymptomatic epilepsy died. In those with symptomatic epilepsy, the mortality risk was 22.9 vs 0.39 expected in a control population (see Ped Neur Briefs June 2001;15:43-44) . Among the 9 deaths in this prospective study, none was sudden unexpected and unexplained (SUDEP). SUDEP risk factors are usually multiple and include early-onset epilepsy, poor seizure control, polytherapy with antiepileptic drugs (AED), and frequent dose adjustments or abrupt AED withdrawal. 
An accurate initial diagnosis is important in counseling parents regarding risk of mortality in childhood epilepsy. Children with secondarily generalized epilepsy associated with functional neurological deficit are particularly at risk of a poor prognosis. Apart from the rare exception of SUDEP in children with tuberous sclerosis, deaths in epilepsy are usually explained.